November 2011  << Back  

  Can't open the PDF? Click here for help.



An open-label trial of acamprosate in the treatment of pathological gambling

Donald W. Black, MD

Department of Psychiatry, University of Iowa, Roy J. and Lucille A. Carver College of Medicine, Iowa City, IA, USA

Dennis P. McNeilly, PsyD

Department of Psychiatry, University of Nebraska Medical Center, Omaha, NE, USA

William J. Burke, MD

Department of Psychiatry, University of Nebraska Medical Center, Omaha, NE, USA

Martha C. Shaw, BS

Department of Psychiatry, University of Iowa, Roy J. and Lucille A. Carver College of Medicine, Iowa City, IA, USA

Jeff Allen, PhD

Department of Psychiatry, University of Iowa, Roy J. and Lucille A. Carver College of Medicine, Iowa City, IA, USA

BACKGROUND: The efficacy and tolerability of acamprosate has been tested in the treatment of pathological gambling (PG). Acamprosate is known to reduce alcohol craving and use in persons with alcohol dependence, and it has been hypothesized that the drug would have a similar effect in individuals with PG.

METHODS: Participants with DSM-IV criteria for PG received acamprosate in an 8-week, open-label trial following a 2-week observation. The primary efficacy measure was the Yale-Brown Obsessive Compulsive Scale modified for PG (Y-BOCS-PG). Secondary efficacy measures included the Gambling Severity Assessment Scale (GSAS), the Clinical Global Impression (CGI) Improvement and Severity Scales, a patient self-rated global rating, the Hamilton Depression Rating Scale (HDRS), the Sheehan Disability Scale (SDS), and the Timeline Follow Back (TLFB). The study was conducted at 2 sites.

RESULTS: Twenty-six participants (11 men, 15 women) had at least 1 post-baseline visit and were included in the analysis. Twenty participants (77%) completed the protocol. Significant improvement was observed in Y-BOCS-PG and GSAS scores, both CGI scales, a patient self-rated global scale, all 3 SDS subscales, and number of gambling episodes. Seventeen participants (65%) were considered responders (ie, achieved “much” or “very much” improvement). Improvements on the HDRS, in money wagered, and in time spent gambling were not significant. Few adverse events were reported.

CONCLUSIONS: The results suggest that acamprosate is well tolerated and may be effective in the treatment of PG.

KEYWORDS: acamprosate, pathological gambling, impulse-control disorders



Pathological gambling (PG) is increasingly being recognized as a public health concern. In 1997, President William J. Clinton established the National Gambling Impact Study Commission to examine the social and economic consequences of gambling. The first such study by a national scientific organization was conducted. In the Commission’s final report, it was estimated that the annual cost for problem and pathological gamblers is $5 billion, with an additional $40 billion in lifetime costs for lost productivity, social services, and creditor losses.1 The effects of PG on families, marriages, and children can be immense, including domestic violence and homelessness.2 It has been estimated that each problem gambler negatively affects 10 to 17 individuals, including family, friends, employers, and the government.3 Pathological gamblers also have high rates of comorbid psychiatric disorders, crime, and suicide.4,5

Despite the growing availability of gambling and its negative impact, PG remains underdiagnosed and undertreated.6 There are currently no FDA-approved medications to treat PG. For several decades, treatment has consisted almost exclusively of participation in the 12-step program Gamblers Anonymous. More recently, cognitive-behavioral therapy and motivational interviewing have become established treatments.7,8

Pharmacologic treatment studies for PG also have recently gained momentum, although results have been mixed. Research has focused largely on 3 drug classes—opiate antagonists, antidepressants, and mood stabilizers. Naltrexone9 and nalmefene,10 both opiate antagonists, were found to be superior to placebo in well-conducted, randomized-controlled trials (RCTs). Among antidepressants, paroxetine was found to be superior to placebo in 1 small RCT,11 but not in a larger multicenter RCT.12 One RCT of fluvoxamine13 produced positive results, while another was negative.14 Open-label trials of bupropion,15 nefazodone,16 citalopram,17 and escitalopram18 have been encouraging. We completed an RCT of bupropion vs placebo19 that was negative, although a comparison of bupropion to naltrexone20 produced similar rates of improvement. There are fewer studies on the use of mood stabilizers in PG. One RCT found lithium to be superior to placebo for pathological gamblers with bipolar spectrum disorders.21 However, RCTs comparing 2 drugs (eg, lithium to valproate22 and topiramate to fluvoxamine23) found similar rates of improvement between the comparison drugs. We recently completed an open-label trial of extended-release carbamazepine, which produced largely positive results.24

The current study investigates the efficacy and tolerability of acamprosate in treating PG. Although the mechanism of action of acamprosate is unknown, it is reported to block glutamate receptors and inhibit hyperactive glutamate signaling.25 Acamprosate was approved by the FDA to treat alcohol dependence, and it appears to reduce cravings. PG has been conceptualized as an addiction similar to alcoholism, with the same tolerance, cravings, and withdrawal.26,27 Individuals with PG often are preoccupied and submissive to their urges and cravings to gamble. We hypothesized that acamprosate would reduce gambling cravings or urges, thereby reducing gambling behavior.


The study took place at the University of Iowa and the University of Nebraska, with each site following an identical study protocol. Data from both sites were pooled for the analysis. Participants were recruited through newspaper or television advertisements and word of mouth.

A total of 39 participants were screened by telephone and 3 did not meet study criteria. Of the 36 evaluated at the screening visit (visit 1), 2 failed to meet study criteria and were excluded; 6 did not return for the baseline visit because of follow-up loss or choosing to discontinue participation. Of the remaining 28 participants, 1 discontinued and another was lost to follow-up after the baseline visit. Therefore, the final sample consisted of 26 participants who completed at least 1 post-baseline visit.

All 11 men and 15 women met DSM-IV-TR28 criteria for PG and were healthy and not depressed. The diagnosis of PG was confirmed through administration of the National Opinion Research Center at the University of Chicago DSM-IV Screen for Gambling Problems (NODS), a structured instrument that assesses current and lifetime PG and has excellent psychometric properties.1 Participants also had to achieve a score of ≥5 on the South Oaks Gambling Screen (SOGS), an instrument considered to be the criterion standard for detecting PG.29

Participants also were administered the Structured Clinical Interview for DSM-IV (SCID)30 to assess Axis I comorbidity. Axis II was screened for using the SCID-Personality.31 All participants were age ≥18, had PG for a minimum of 1 year, spoke adequate English, and had ≥2 gambling episodes within the 2-week screening period. The study was approved by each of the 2 universities’ institutional review boards. Participants gave written informed consent at their screening visit.

We excluded those with significant or unstable medical disorders; a history of schizophrenia, schizoaffective disorder, or bipolar mood disorder; severe borderline or antisocial personality disorder; current drug or alcohol dependence; an organic mental disorder; and those who were deemed to be at risk for suicide by the study clinician. Participants were excluded if they had a Hamilton Depression Rating Scale (HDRS)32 score ≥18 at their baseline visit (visit 2) or met DSM-IV criteria for major depressive disorder. Participants were not eligible if they had received acamprosate in the past, or had entered individual or group psychotherapy for the treatment of PG during the 3 months before entering the study. Participants had to be free of all psychotropic medications for at least 2 weeks prior to their baseline visit (3 weeks for monoamine oxidase inhibitors, 4 weeks for fluoxetine, and 3 months for long-acting phenothiazines).

The primary efficacy measure was the Yale-Brown Obsessive Compulsive Scale modified for PG (Y-BOCS-PG).33 Secondary efficacy measures included the Gambling Severity Assessment Scale (GSAS)9; the Clinical Global Impression (CGI) Severity (S) and Improvement (I) Scales34; a self-rated patient global rating patterned after the CGI Improvement scale; the HDRS31; the Sheehan Disability Scale (SDS)35; the Attention-Deficit/Hyperactivity Disorder (ADHD) Checklist36; and the Timeline Follow Back (TLFB),37 an instrument used to track gambling frequency, time spent, and money wagered. The ADHD Checklist was used as a secondary outcome measure because of the frequent comorbidity of PG and ADHD, and because it is a measure of impulsivity.4

Participants who met study criteria entered a 2-week observation period before starting the drug. At visit 2, all participants were started on acamprosate, 1,998 mg divided into 3 equal doses. The study psychiatrist and study coordinator limited their interaction to discussion of clinical history, review of recent gambling behavior, medication adverse effects, and to providing general support. No advice was given as to how participants might reduce their gambling behavior.

Safety assessments were based on vital signs, clinical laboratory tests including biochemistry and hematology, and ECG. Vital signs were assessed at each visit, and the other tests were obtained at the initial screening and at week 8, the final visit. Adverse events were recorded by the study psychiatrist at each visit and were spontaneously reported by the participants in response to non-leading questions.

The linear-mixed effects model, which characterizes each participant’s measures with a baseline value (intercept) and rate of change value (slope), was used to test the efficacy of acamprosate. We hypothesized that the mean slope for the primary outcome variable (Y-BOCS-PG) would be <0, indicating a positive treatment effect. The linear-mixed effects model was applied to all semi-continuous efficacy measures, including the GSAS; the CGI-I and CGI-S scales; the patient self-rated global rating; the HDRS; the SDS subscales for work/school, social, and family; and the ADHD Checklist. The TLFB outcomes (money wagered, time spent gambling, and number of episodes) appeared to be highly skewed. For example, the median dollars wagered per day was $121 at baseline, but the mean was $262. To correct for this, and to satisfy the normality assumption of the statistical model, we transformed money wagered (in dollars), time spent gambling (in minutes), and number of gambling episodes, to the natural logarithm scale. We then applied the same linear mixed-effect model to test if the mean rate of change of these measures was 0. For all analyses, treatment site was included as a covariate to account for outcome differences associated with the 2 university settings. In a secondary analysis, we extended the linear mixed-effect models by including a treatment site by slope interaction; this allowed us to test whether the treatment effect varied significantly across the 2 sites. In another secondary analysis, we included substance abuse history by slope interaction; this allowed us to test whether the treatment effect was different for patients with a substance abuse history. Statistical analyses were carried out using SAS version 9.1 (SAS Institute Inc., Cary, NC).38 Linear mixed-effect models were fit with the MIXED procedure using a 2-tailed significance level of .05.


Of the 26 participants who entered the study, 20 (77%) completed the entire 8-week protocol. Six participants did not complete the study; 1 discontinued after 2 weeks of treatment, another left after 4 weeks of treatment, and 4 left after 6 weeks of treatment. Reasons for non-completion include moving to another town (n = 2) and being lost to follow-up (n = 3). One participant was removed from the study after becoming suicidal.

Demographic and illness characteristics are presented in TABLE 1. The mean (standard deviation [SD]) age of the 26 participants was 48.0 (16.4); 11 (42%) were men and 15 (58%) were women, 8 (31%) were married, 8 (31%) were divorced or separated, 7 (27%) were never married, and 3 (12%) were widowed. Most (81%) were white; 4 (15%) were African American, and 1 was Hispanic. The mean (SD) age at PG onset was 40.2 (13.7), and the mean (SD) duration of illness was 7.8 (7.4) years (range, 1 to 31 years). The mean (SD) duration of education was 14.0 (1.7) years, with 8 of the participants (31%) having earned a bachelor’s degree. Nine of the participants (35%) had prior treatment for PG. Seven (27%) had a lifetime substance use disorder, 8 (31%) had a lifetime mood disorder, and 7 (27%) had a lifetime anxiety disorder.


Demographic and clinical characteristics of 26 participants with PG

Variable Mean (SD)
Age, years 48.0 (16.4)
Age at PG onset, years 40.2 (13.7)
Duration of PG, years 7.8 (7.4)
Education, years 14.0 (1.7)
NODS 16.2 (4.0)
SOGS 13.2 (3.4)
Sex N (%)
Men 11 (42%)
Women 15 (58%)
Marital status
Married 8 (31%)
Divorced/separated 8 (31%)
Never married 7 (27%)
Widowed 3 (12%)
White 21 (81%)
African American 4 (15%)
Hispanic 1 (4%)
Prior PG treatment
Yes 9 (35%)
No 16 (62%)
Unknown 1 (4%)
NODS: National Opinion Research Center at the University of Chicago DSM-IV Screen for Gambling Problems; PG: pathological gambling; SOGS: South Oaks Gambling Screen.

TABLE 2 provides a summary of baseline scores and statistical test results on the primary and secondary efficacy measures. The mean scores of the baseline measures indicate that patients had moderate to severe PG. At baseline, 22 of the 26 participants were gambling at least weekly and 18 (69%) gambled at least 3 times per week. Improvement was noted on all efficacy scales, and 17 of 26 participants (65%) met the a priori definition for our responder status.

Over the 8 weeks, there was significant improvement in most efficacy measures (TABLE 2), including the Y-BOCS-PG total score and both subscales. The average participant’s Y-BOCS-PG total score dropped from 21.0 at baseline to 10.5 by the final visit (FIGURE 1). The participants also improved significantly with respect to the GSAS, CGI, patient self-ratings of improvement, all 3 SDS subscales, and ADHD Checklist. The improvement observed on the HDRS was not significant. For the TLFB measures (money wagered, time spent gambling, and number of episodes per week), we observed significant improvement in number of gambling episodes. We also observed an average weekly decrease of .08 in log number of gambling episodes, which translates to an 8% weekly decrease in number of gambling episodes. This degree of improvement is both statistically significant and clinically meaningful.

Improvement from baseline to study end in gambling severity (GSAS) differed across the 2 sites: patients from the Nebraska site improved more than patients from Iowa (P = .023). Severity at baseline was different across sites for the CGI-S and Y-BOCS urge/thought subscale. In each case, patients from Iowa were more severe initially (P = .037, P = .048, respectively). Improvement from baseline to study end on the SDS work/school subscale was greater for participants with a substance abuse history (P = .047). Substance abuse history was not related to initial severity in any outcomes.

Acamprosate was found to be well tolerated. Thirteen participants reported diarrhea, 6 reported somnolence, 3 reported cold-like symptoms, and 2 reported flatulence. These adverse effects were all of mild severity. As noted earlier, 1 participant reported suicidal ideation that was not thought to be related to the drug and this patient was removed from the study. There were no changes in laboratory parameters including hematology, serum biochemistries, and ECG.


Statistical tests of efficacy measures in 19 participants with PG

Efficacy measure Baseline mean (SD) Mean slope (SE) Test of mean slope = 0
F df P value
  Total 21.2 (5.2) –.985 (.158) 39.4 1,25 <.001
  Urge/thought subscale 10.7 (2.4) –.522 (.076) 47.5 1,25 <.001
  Behavior subscale 10.5 (3.2) –.465 (.087) 28.5 1,25 <.001
GSAS 26.7 (8.1) –1.117 (.228) 24.0 1,25 <.001
CGI-I NA –.145 (.031) 21.4 1,25 <.001
CGI-S 5.2 (1.0) –.199 (.035) 32.6 1,25 <.001
  Self-Rated Global Improvement NA –.116 (.032) 13.2 1,25 .001
HDRS (17-item) 3.9 (3.3) –.078 (.070) 1.3 1,24 .273
  Work/school 2.7 (2.8) –.206 (.058) 12.8 1,25 .002
  Social 3.1 (2.5) –.271 (.052) 27.2 1,25 <.001
  Family 4.0 (2.7) –.301 (.054) 31.7 1,25 <.001
ADHD Checklist 8.0 (7.7) –.339 (.106) 10.2 1,25 .004
  Wagers per week (log dollars) 4.4 (2.4) –.055 (.062) 0.8 1,25 .384
  Time spent per week (log minutes) 4.7 (2.4) –.071 (.066) 1.2 1,25 .289
  Episodes per week (log of count) 1.0 (1.1) –.080 (.019) 17.4 1,25 <.001
ADHD: attention-deficit/hyperactivity disorder; CGI: Clinical Global Impression; GSAS: Gambling Severity Assessment Scale; HDRS: Hamilton Depression Rating Scale; PG: pathological gambling; SD: standard deviation; SDS: Sheehan Disability Scale; SE: standard error; TLFB: Timeline Follow Back; Y-BOCS-PG: Yale-Brown Obsessive Compulsive Scale modified for Pathological Gambling.

FIGURE 1: Changes in Y-BOCS-PG mean scores

Y-BOCS-PG: Yale-Brown Obsessive Compulsive Scale modified for Pathological Gambling.


The results suggest that acamprosate may be effective in the treatment of PG. In addition to its therapeutic benefit, the drug was well tolerated. All of the primary and secondary outcome variables improved except for the HDRS and 2 subscales of the TLFB. The former is not surprising because acamprosate is not an antidepressant, and participants had relatively low levels of depressive symptoms at baseline. The latter findings suggest that gambling improvement has several important aspects, and that global improvement may not necessarily depend on measurable improvements in wagers or time spent per week. Functional status improved, as indicated by the changes observed in the 3 SDS subscales assessing work/school, social, and family functioning. Although no participants had a clinical diagnosis of ADHD, improvement was seen in their ADHD Checklist scores, suggesting that a reduction in gambling behavior may be accompanied by a more general improvement in impulsivity, one of the main components of ADHD.

Most participants improved in terms of their gambling urges and behaviors, yet only 35% were abstinent in the final 2 weeks of study participation, indicating that additional research in this area is warranted. Improvement without abstinence may not be viewed as acceptable by many in the gambling treatment community. Nonetheless, there has been growing interest in the concept of “controlled gambling.”39

There are several limitations to acknowledge. First, the sample was relatively small, and there was no comparison group; therefore, our results could have resulted from the placebo effect. An RCT likely would help differentiate drug- from placebo-related improvement. In addition, non-blind assessments could have inadvertently complicated the interpretation of study results. Another limitation is the relatively short length of the trial; a longer trial (eg, 12 to 16 weeks) could provide a better test of drug response because the placebo effect tends to fade with time. Furthermore, because there were few minority participants, results cannot be generalized to minority populations. Finally, although reliable and valid assessments were used, clinician-rated instruments may not have been administered in an identical fashion at the 2 sites. That said, the analysis showed that the treatment effect differed across sites only for the GSAS, a self-rated instrument. We did attempt to mitigate the potential problem of having 2 sites by covarying for site effect in our analysis.


We were able to show through a 2-site, open-label study that acamprosate has acceptable safety and tolerability in patients with PG. The drug may be effective in treating PG because most patients improved in primary and secondary efficacy measures. It also produced “real-world” improvements in functional status. We believe an appropriately powered randomized controlled trial should be organized to provide a proper test of acamprosate in this population.

ACKNOWLEDGEMENT: This work was supported in part through an investigator-initiated grant from Forest Research Institute/Forest Laboratories.

DISCLOSURES: Dr. Black receives grant/research support from Psyadon Pharmaceuticals and AstraZeneca and receives royalties from American Psychiatric Publishing, Inc., and Oxford University Press. Drs. McNeilly, Burke, and Allen and Ms. Shaw report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.


  1.  Gambling Impact and Behavior Study.  Chicago IL: National Opinion Research Center at the University of Chicago (NORC); 1999.
  2. Shaw MC, Forbush KT, Schlinder J, et al. The effect of pathological gambling on families, marriages, and children. CNS Spectr. 2007;12:615–622.
  3. Politzer R, Yesalis C, Hudak C. The epidemiologic model and the risk of legalized gambling: where are we headed? Health Values. 1992;16:20–27.
  4. Black DW, Shaw MC. Psychiatric comorbidity associated with pathological gambling. Psychiatric Times. 2008;25:14–18.
  5. Shaffer HJ, Korn DA. Gambling and related mental disorders: a public health analysis. Annu Rev Public Health. 2002;23:171–212.
  6. Cunningham JA. Little use of treatment among problem gamblers. Psychiatr Serv. 2005;56:1024–1025.
  7. Ladouceur R, Sylvain C, Boutin C, et al. Cognitive treatment of pathological gambling. J Nerv Ment Dis. 2001;189:774–780.
  8. Petry NM. Psychosocial treatments for pathological gamblers. Psychiatric Annals. 2002;32:192–196.
  9. Kim SW, Grant JE, Adson DE, et al. Double-blind naltrexone and placebo comparison study in the treatment of pathological gambling. Biol Psychiatry. 2001;49:914–921.
  10. Grant JE, Potenza MN, Hollander E, et al. Multicenter investigation of the opioid antagonist nalmefene in the treatment of pathological gambling. Am J Psychiatry. 2006;163:303–312.
  11. Kim SW, Grant JE, Adson DE, et al. A double-blind placebo-controlled study of the efficacy and safety of paroxetine in the treatment of pathological gambling. J Clin Psychiatry. 2002;63:501–507.
  12. Grant JE, Kim SW, Potenza MN, et al. Paroxetine treatment of pathological gambling: a multi-centre randomized controlled trial. Int Clin Psychopharmacol. 2003;18:243–249.
  13. Hollander E, DeCaria CM, Finkell JN, et al. A randomized double-blind fluvoxamine/placebo crossover trial in pathologic gambling. Biol Psychiatry. 2000;47:813–817.
  14. Blanco C, Petkova E, Ibáñez A, et al. A pilot placebo-controlled study of fluvoxamine for pathological gambling. Ann Clin Psychiatry. 2002;14:9–15.
  15. Black DW. An open-label trial of bupropion in the treatment of pathologic gambling. J Clin Psychopharmacol. 2004;24:108–110.
  16. Pallanti S, Baldini Rossi N, Sood E, et al. Nefazodone treatment of pathological gambling: a prospective open-label controlled trial. J Clin Psychiatry. 2002; 63:1034–1039.
  17. Zimmerman M, Breen RB, Posternak MA. An open-label study of citalopram in the treatment of pathological gambling. J Clin Psychiatry. 2002;63:44–48.
  18. Black DW, Shaw M, Forbush KT, et al. An open-label trial of escitalopram in the treatment of pathological gambling. Clin Neuropharmacol. 2007;30:206–212.
  19. Black DW, Arndt S, Coryell WH, et al. Bupropion in the treatment of pathological gambling: a randomized, double-blind, placebo-controlled, flexible-dose study. J Clin Psychopharmacol. 2007;27:143–150.
  20. Dannon PN, Lowengrub K, Musin E, et al. Sustained-release bupropion versus naltrexone in the treatment of pathological gambling: a preliminary blind-rater study. J Clin Psychopharmacol. 2005;25:593–596.
  21. Hollander E, Pallanti S, Allen A, et al. Does sustained-release lithium reduce impulsive gambling and affective instability versus placebo in pathological gamblers with bipolar spectrum disorders? Am J Psychiatry. 2005;162:137–145.
  22. Pallanti S, Quercioli L, Sood E, et al. Lithium and valproate treatment of pathological gambling: a randomized single-blind study. J Clin Psychiatry. 2002;63:559–564.
  23. Dannon PN, Lowengrub K, Gonopolski Y, et al. Topiramate versus fluvoxamine in the treatment of pathological gambling: a randomized, blind-rater comparison study. Clin Neuropharmacol. 2005;28:6–10.
  24. Black DW, Shaw MC, Allen J. Extended release carbamazepine in the treatment of pathological gambling: an open-label study. Prog Neuropsychopharmacol Biol Psychiatry. 2008;32:1191–1194.
  25. De Witte P, Littleton J, Parot P, et al. Neuroprotective and abstinence-promoting effects of acamprosate: elucidating the mechanism of action. CNS Drugs. 2005;19:517–537.
  26. Shaffer HJ, LaPlante DA, LaBrie RA, et al. Toward a syndrome model of addiction: multiple expressions, common etiology. Harv Rev Psychiatry. 2004;12:367–374.
  27. Wray I, Dickerson MG. Cessation of high frequency gambling and “withdrawal” symptoms. Br J Addict. 1981;76:401–405.
  28.  Diagnostic and statistical manual of mental disorders, 4th ed, text rev Washington, DC: American Psychiatric Association; 2000.
  29. Lesieur HR, Blume SB. The South Oaks Gambling Screen (SOGS): a new instrument for the identification of pathological gamblers. Am J Psychiatry. 1987;144:1184–1188.
  30. First MB, Spitzer RL, Gibbon M, et al. Structured Clinical Interview for Axis I DSM-IV Disorders – Patient Edition (SCID-I/P). New York: Biometrics Research Department, NY State Psychiatric Institute: 1995.
  31. Pfohl B, Zimmerman M, Blum N. A Structured Interview for DSM-IV Personality (SIDP-IV). Washington, DC: American Psychiatric Press; 1997.
  32. Hamilton M. Development of a rating scale for primary depressive illness. Br J Soc Clin Psychol. 1967;6:278–296.
  33. Pallanti S, DeCaria CM, Grant JE, et al. Reliability and validity of the pathological gambling adaptation of the Yale-Brown Obsessive-Compulsive Scale (PG-YBOCS). J Gambl Stud. 2005;21:431–443.
  34. Guy W. ECDEU assessment manual for psychopharmacology—Revised (DHEW Publ No ADM 76-338). Rockville MD: U.S. Department of Health, Education, and Welfare, Public Health Service, Alcohol, Drug Abuse, and Mental Health Administration, NIMH Psychopharmacology Research Branch, Division of Extramural Research Programs; 1976.
  35. Sheehan DV. The anxiety disease. New York NY: Bantam Books, Inc.; 1986.
  36. Barkley RA. Attention-deficit hyperactivity disorder: handbook for diagnosis and treatment. New York NY: Guilford Press; 1990.
  37. Sobell LC. Alcohol timeline followback user’s manual. Toronto Canada: Addiction Research Foundation; 1995.
  38.  SAS Institute Inc., SAS/STAT. 9.1 Users Guide. Cary, NC: SAS Institute, Inc.; 2004.
  39. Blaszczynski A, McConaghy N, Frankova A. Control versus abstinence in the treatment of pathological gambling: a two to nine year follow-up. Br J Addict. 1991;86:299–306.

CORRESPONDENCE: Donald W. Black, MD, Psychiatry Research, 2-126B MEB, University of Iowa, Carver College of Medicine, Iowa City, IA 52242 USA, E-MAIL: